Light-harvesting antennas in the photosynthesis systems of plants are known to be high-efficiency photon collectors. Here we use light-harvesting antennas in dye-sensitized solar cells.1 The concept of spectral sensitization of ZnO by light-harvesting antennas composed of multiple stacked dyes is shown in Fig. 1. Dyes with different energy gaps are stacked in a designated sequence so that a wide range of wavelengths are absorbed. The resulting broad absorption spectrum is a superposition of the narrow absorption bands of the individual dye molecules. This energy scheme is expected to decrease the heat generation caused by excess photon energy in the light absorption process.2 We have previously sensitized ZnO using two-dye stacked structures consisting of a p-type dye and an n-type dye to widen the photocurrent spectral width.2 In the present work, we investigate sensitization of ZnO using three-dye stacked light-harvesting antennas grown by liquid-phase molecular layer deposition (LP-MLD)2-4using n- and p-type dye molecules. Meier5reported that p-type dyes such as rose bengal (RB), eosin (EO) and fluorescein (FL) tend to have negative charge because they accept electrons, while n-type dyes like crystal violet (CV) and brilliant green (BG) typically have positive charge as they donate electrons. Because ZnO is an n-type material, as shown in Fig. 2, three-dye stacked structures can be grown using a single dye in each step by providing source molecules with a sequence of p-, n-, and p-type. Here, the LP-MLD process was achieved by electrostatic force. Namely, the attraction between p–n and repulsion between p–p and n–n molecules resulted in self-limiting growth, allowing us to construct structures of [n-type ZnO/p-type dye/n-type dye/p-type dye]. Samples for photocurrent measurements were prepared as follows. In step 1, after a ZnO powder layer (~5 μm thick) was formed on a glass substrate with a slit-type indium tin oxide (ITO) electrode (slit width of 60 μm), the layer was immersed in an isopropyl alcohol (IPA) solution of the first p-type dye (p1). In step 2, the layer was rinsed with IPA, and immersed in an IPA solution of n-type dye (n1) before the layer was dried. In step 3, the layer was immersed in an IPA solution of the second p-type dye (p2) in the same manner. This procedure gave three-dye stacked light-harvesting antennas with a structure of [ZnO/p1/n1/p2]. Photocurrent spectra were then measured by applying a 3-V potential across the slit-type ITO electrodes over a wavelength range of 400–700 nm. Fig. 3 shows reflection spectra of ZnO powder layers with adsorbed RB, EO, FL, CV, and BG dyes. By stacking these five dyes, photocurrent spectra were widened to cover the wavelength range from ~450 to 700 nm. In Fig. 4(a), photocurrent spectra measured following each growth step are shown for a light-harvesting antenna of [ZnO/EO/CV/RB]. When EO was deposited on a ZnO surface during Step 1, large photocurrents were generated over the wavelength region from ~500 to ~550 nm. When CV was deposited on EO in Step 2, photocurrents caused by CV were generated in the wavelength region longer than ~550 nm in addition to the photocurrent from EO absorption. When RB was deposited on CV during Step 3, photocurrents caused by RB were observed around ~570 nm. These results suggest that the multi-dye stacked light-harvesting antenna is effective to widen the wavelength region for sensitization of ZnO. It is important to arrange the dye molecules so that an energy-level scheme with a gradual energy slope similar to that of light-harvesting antennas in plants is formed. This facilitates electron transfer from the multi-dye stacked structure to the semiconductor to achieve high-efficiency sensitization. Fig. 4(b) shows an energy-level scheme determined from photoemission yield spectroscopy in air (PYSA)6measurements for [ZnO/EO/CV/RB]. The direction of the energy slope is opposite to the preferred direction. As a future challenge, to further improve sensitization efficiency, the energy-level scheme of the dye molecules should be optimized. References 1 B. O’Regan and M. Gratzel, Nature 353, 737 (1991). 2 T. Yoshimura, H. Watanabe, and C. Yoshino, J. Electrochem. Soc., 58, 51 (2011). 3T. Yoshimura, Japanese Patent, Tokukai Hei3-60487 (1991) [in Japanese]. 4T. Yoshimura, Japanese Patent, Tokukai 2009-60487 (2009) [in Japanese]. 5 H. Meier, J. Phys. Chem. 69, 719 (1965). 6 Y. Nakajima, D. Yamashita, A. Ishizaki, B. Pellissier, and M. Uda, Mater.Res. Symp. Proc. 1029, 1029-F04-02 (2008). Figure 1
Read full abstract